HR: 14:40h
AN: S32D-05 [PDF]
TI: Bayesian Analysis of Aftershock Sequences
AU: * Chen, Y
EM: ychen@stat.ncu.edu.tw
AF: Institute of Statistics, National Central University, #300 Junda Road, Chung-Li, 32054
Taiwan
AU: Fan, T
EM: thfan@stat.ncu.edu.tw
AF: Institute of Statistics, National Central University, #300 Junda Road, Chung-Li, 32054
Taiwan
AU: Liu, J
EM: jyliu@jupiter.ss.ncu.edu.tw
AF: Institute of Space Science, National Central University, #300 Junda Road, Chung-Li, 32054
Taiwan
AB:
The Reasenberg-Jones (1989) or RJ model describes the time-magnitude hazard of an aftershock sequence by combining the
modified OmoriÝs law (Utsu, 1961) and the Gutenberg-Richter (1954) rule. The parameters estimated separately from the
power-decaying law and magnitude-frequency relation are then integrated for analyzing aftershock sequences in California.
Wiemer (2000) further applied the RJ model to spatial grids over a possible aftershock area for evaluating the hazard or
probability of a future aftershock. In this study, we propose a Bayesian analysis of the RJ model for a aftershock sequence
with a non-informative prior or prior distributions based on previous aftershock sequences. Note that, under a
non-informative prior, the Bayesian version of the RJ model based on posterior modes of the related parameters provides with
an approximation to the estimated RJ model which maximizes the associated likelihood. The three versions of the RJ model,
including the Bayesain, maximum likelihood and the integrated ones, will be applied to the Chi-Chi (M=7.6, 1999/9/20 UT)
aftershocks for comparative purpose.
DE: 7223 Seismic hazard assessment and prediction
SC: Seismology [S]
MN: 2003 Fall Meeting